A review of 1,841 Vaccine Adverse Event. Reporting System reports, submitted between 1998 and 2001 for adverse events associated with anthrax vaccine ...
Military Hospitalizations Among Deployed US Service Members Following Anthrax Vaccination, 1998–2001
T. S. Wells P. A. Sato T. C. Smith L. Z. Wang R. J. Reed M. A.K. Ryan
Naval Health Research Center Report No. 04-23
. Approved for public release: distribution is unlimited.
Naval Health Research Center 140 Sylvester Road San Diego, California 92106
[Human Vaccines 2:2. 54-59. March/April 2006]; ©2006 Landes Bioscience
Research Paper
Military Hospitalizations Among Deployed US Service Members Following Anthrax Vaccination, 1998-2001 Timothy Steven Wells* Paul A. Sato Tyler Clain Smith Linda Zhenling Wang Robert John Reed Margaret Angela Kappel Ryan Depart...nl of Defense ClInter for Deploymenl Health Reseorlh; Novo! Health Researth Cenler; Son Diego, Colifomia USA 'CorT8spondence 10: TImothy S. Wells; Biome
ACKNOWLEDGEMENTS The authors thank Mr. Scott L. Seggerman and Mr. Jack White of the Defense Manpower Data Center. Seaside, CA. for their assistance in procuring the demographic, vaccination, and military-specific data. We are also extremely grateful to the following professional collaborators: Neal A. Halsey, M.D., Department of International Health. The lohns Hopkins Bloomberg School of Public Health; Dr. Brian P. Murphy. Naval Health Research Center; Dr. Phillip R. Pittman, M.D.. M.P.H.. Chief, Division of Medicine, Emeritus, US Army Medical Research Institure of Infectious Diseases; Dr. Gregory C. Gray. College of Public Health, University of Iowa; Dr. John D. Grabenstein, Military Vaccine Agency, US Army Medical Command; and Dr. Ruger L. Gibson, Executive Secretary. Armed Forces Epidemiological Board. Dr. Wells had full access to all of the data in the study and takes responsibiJiry for rhe integrity of the data and the accuracy of the data analysis.
54
ABSTRACT Safety concerns have confronted the Department of Defense Anthrax Vaccine Immunization Program since inception in 1998. To determine if anthrax vaccination was associated with an increased risk of hospitalization, a historical cohort study utilizing preand post-anthrax-vaccination hospitalizations was undertaken and analyzed with Cox proportional hazards models. The study population consisted of 170,723 active duty US service members who were anthrax-vaccinated and deployed during the time period January 1, 1998 to December 31, 2001 . Study outcomes included hospitalizations due to any-eause, 14 broad International Classification of Diseases diagnostic categories, autoimmune organ specific and organ non-specific hospitalizations, and asthma. After adjustment, anthrax vaccination was associated with significantly fewer hospitalizations for any-eause, diseases of the blood and blood forming organs, and diseases of the respiratory system. Comparing anthrax post-vaccination hospitalization experience with the pre-vaccination period resulted in no signiticant increased hazard for any of the hospitalization outcomes studied. Although there was no apparent increase in risk of morbidity in this study population, the relationship between anthrax vaccine and deployment on health outcomes among US service memoers needs further study. The history and rationale behind the development of the Department of Defense (DoD) Anthrax Vaccine Immunization Program has been previously documented. 1 Since inception, this vaccine program has been surrounded by controversy, including service members who have expressed health concerns over receiving the vaccine, lapses in vaccine production, and court-ordered injunctions. 2- 5 Between May 2000 and September 2000 the US General Accounting Office conducted a stratified random survey of 1,253 Air National Guard and Air Force Reserve personnel who were currently serving or had separated. 6 Sixty-five percent of subjects responded and 41 % were skeptical of the DoD biological warfare threat assessment. More than 60% were concerned about short-term and long-term health events associated with anthrax vaccination. Some researchers and members of the media have hypothesized a causal link between the anthrax vaccine and symptoms related to Gulf War serviceJ,8 Case reports have identified temporal associations between anthrax vaccination and delayed-type hypersensitivity reaction,9 hypersensitiVity pneumonitis,IO optic neuritis, 11 and pemphigus vulgaris. 12 In response to growing concerns regarding the safety of the anthrax vaccine, a number of epidemiological studies were conducted. A review of 1,841 Vaccine Adverse Event Reporting System reports, submitted between 1998 and 2001 for adverse events associated with anthrax vaccine, found no evidence of an unusually high rate for serious adverse events or other medically important adverse events. 13 A study utilizing pre- and postanthrax vaccination cohorts to calculate adjusted rates for over 843 specific diagnoses found positive associations, but after accounting for multiple comparisons and the potential for misclassifications, it was concluded that anthrax vaccine has few clinically significant adverse effects. 14 Using preliminary (I998) data from this study, Saw et al. assessed the relations between 14 broad categories of hospital discharge diagnoses and anthrax vaccination using Cox proportional hazarcls modeling and found that anthrax-vaccinated, active-duty service members were at equal or less risk of hospitalization. 15 A cohort of anthrax vaccinated active-duty health care professionals was followed 1-2 weeks post-vaccination. They also were assessed via two nested designs that used either self-reported survey data or rates from outpatient visits and hospitalizations. The authors noted that although the vaccine was considerably reactogenic, very few serious adverse events were observed. 16 A study of 1,600 scientists and maintenance workers who had received a total of 10,722 anthrax vaccine doses reported incidence rates of approximately I % and 4% for systemic and local Human Vaccines
2006; Vol. 2 Issue 2
Anthrax Vaccination and Military Hospitalizations
adverse events, respectively.I7 A study of US Air Force service members who had been deployed to Southwest Asia found no increase in ambulatory medical visits among anthrax vaccinees. 18 Similarly, a study of anthrax-vaccinated individuals among the Canadian military found no increased risk for medical visits within eight months of vaccination. 19 A large historical cohort study of active-duty US Army personnel found no relationship to disability evaluations. 2o Finally, the Institute of Medicine reviewed the scientific literature on the safety and efficacy of the anthrax vaccine and determined it was effective and "reasonably safe".21 The objective of our study was to compare the adjusted risk of hospitalization among pre- and post-anthrax-vaccinated and deployed US service members who had served on active duty between 1998 and 2001.
METHODS Data. Methods used to assemble the analytic cohort were previously described. Is A historical cohort design was used to gather elecrronically available demographic, military-specific, and health outcomes dara. Demographic data were obtained ITom the Defense Manpower Dara Center, and anthrax vaccination status was obtained from the Defense Enrollment Eligibility Reporting System. Hospital discharge diagnoses assigning an International Classification of Diseases, Ninth Revision, Clinical Modifications (lCD-9-CM) code were obtained from the Standard Inpatient Data Record and the Health Care Service Record. All files were obtained in electronic format and were linked by personal identifiers.· Study population. All US service members who had served on active duty a minimum of 30 consecutive days or longer between January 1, 1998, and December 31, 2001, were identified and described. However, to account for unmeasured confounding, and to decrease heterogeneity among subjects, the analytic cohort was restricted to those service members who had received one or more anthrax vaccinations and had been deployed to high-threat areas, such as southwest Asia and received one or more pay entitlements. Months deployed was estimated based on the number of monrhs these entitlements were received. Outcomes. The objective of this study was to identify any yet-to-beknown severe adverse health effects, excluding mortality and pregnancy-related outcomes, associated with anthrax vaccination. Outcomes were limited to hospitalizations to focus on severe morbidity and to minimize possible misclassification associated with ambulatory diagnoses or self-reported morbidity. A priori outcomes included any-cause hospitalization, hospitalizations among 14 broad ICD-9-CM categories as described by Sato et al., 15 autoimmune organ-specific or organ-non-specific disease hospitalizations,22 and specific diagnoses thought to possibly manifest ITom vaccine exposure. 9.1 2 Statistical analyses. Pearson chi-square tests and t resrs were used to calculate univariate measures of association for demographic and military-specific variables. Those variables significant at p $ 0.15 were entered into a post-vaccination versus pre-vaccination multivatiable Cox proportional hazards model. Posr-vaccination person-time was calculated by cumulating successive 42-day risk windows following each recorded anthrax dose. For example, if rhe member remained on the correct dosing schedule, he or she would remain under consrant observarion unril 42 days alTer the rhird dose. Post-vaccination person-time began on the day vaccinated and conrinued until whichever of the following occurred first: (1) the date of first admission, (2) the date of separation from the military, (3) 42 days afTer date vaccinated, or (4) the end of the srudy, December 31. 2001. Forty-two day observation periods were chosen as the best compromise between identifying acute outcomes that may be associated with anthrax vaccination yet preventing spurious associations simply due to temporal sequencing. Three previous studies regarding vaccine safety suggest the use of risk periods somewhere between 3D-days and 6_weeks. 23 ·2S Pre-vaccination person-time was set at 130 days prior to first vaccination, which corresponded to the mean length of follow-up for all post-vaccination person-time. To account for the correlation between a subject's pre-vaccination and post-vaccination
www.landesbioscience.com
IllOOOO
C
90000
~
1liOOO
"
60000
i
.~
t
~OOO
